AI evidence extraction

The Evolving Landscape of AI in Education: A Systematic Review of Contemporary Research (2024-2025)

Mohamed Yusuf Adan · SERDEC education journal. · 2026

AI-generated evidence extraction, verified across multiple analytical personas. Not a substitute for the peer-reviewed original.

6/10

Relevance

1/4

Quality (LMQS)

Evidence

Citations

29.21

FWCI

Top 10%

Impact

This is an AI analysis. Read the peer-reviewed original at the publisher: https://doi.org/10.70595/sej115

Methodology & findings

Study design

Systematic review conducted in accordance with PRISMA 2020 guidelines.

Sample

N = 37, 3 groups

Primary method

Narrative synthesis approach (non-quantitative synthesis); descriptive analysis of publication trends, research methods, thematic focus areas, and educational contexts.

Main result

The study found that "the findings reveal a field dominated by meta-research, including bibliometric analyses, conceptual papers, and systematic reviews, with limited primary empirical evidence assessing real world educational outcomes." Major thematic areas include personalized and self-regulated learning, generative AI and assessment, AI literacy, ethical and equity concerns, and institutional governance, with "higher education and professional training contexts are disproportionately represented, while school-level education receives comparatively limited attention."

Reports effect sizes.

Research paradigm

Critical realism / pragmatism (systematic review synthesizing diverse empirical and theoretical work)

Author conclusions

The authors conclude that "the study provides a timely synthesis of contemporary AI-in-education research, identifying critical evidence gaps and governance challenges. It underscores the need for empirically grounded, ethically informed, and context-sensitive approaches to AI integration that move beyond conceptual promise toward sustainable and equitable educational practice."

Risk of bias

Publication bias (only peer-reviewed studies included; gray literature excluded); Language bias (search likely limited to English-language databases); Database coverage bias (4 databases searched; potential omission of studies in other databases); Selection bias (eligibility criteria for the 37 studies not explicitly detailed in abstract); Temporal bias (2-year window may not capture full landscape); Potential language bias (English-language searches only, not explicitly stated but implied by databases used). Publication bias toward meta-research and conceptual studies. Geographic and institutional bias favoring higher education over school-level contexts. Time-limited search window (2024-2025 only) may miss emerging trends.; Publication bias (only peer-reviewed studies included, likely favoring positive results); Database selection bias (limited to academic databases, may exclude grey literature); Language bias (not specified if non-English papers were included); Time period limitation (2024-2025 only, may not capture foundational research)

Limitations

The paper states: "The findings reveal a field dominated by meta-research, including bibliometric analyses, conceptual papers, and systematic reviews, with limited primary empirical evidence assessing real world educational outcomes." Additionally, "the review highlights a growing research–practice gap, as technological adoption outpaces empirical validation and policy development." The narrow publication window (2024-2025) and restriction to peer-reviewed sources may limit comprehensiveness.

Open questions raised

Limited primary empirical evidence assessing real-world educational outcomes
Underrepresentation of school-level education research (higher education and professional training disproportionately represented)
Growing research-practice gap: technological adoption outpaces empirical validation and policy development
Need for empirically grounded, ethically informed, and context-sensitive approaches to AI integration
Critical evidence gaps and governance challenges requiring further investigation
Critical evidence gaps include: limited primary empirical evidence on real-world educational outcomes; need for school-level education research; need for empirically grounded approaches to AI integration; gaps in policy development; need for ethically informed and context-sensitive implementation frameworks; underrepresentation of equity and access considerations in practical deployment.

Data: not_statedCode: not_statedExtracted from: pdfAgreement 65%

Explore related topics

Institutional Policy AI Governance